Most of my Cyber Defense 1 students finish guided labs cleanly and still stall when a Linux host "is behaving oddly" and the next command is not on the page. The progressive project for IDET 6315 is that stall, narrowed to one unit: Challenge-Based Defense Scenarios on the Open Cyber Range. The learners are working adults in a hybrid section. They do a lot of the work alone, on a laptop over WireGuard, in an evening block that rarely gives them more than an hour.
What North forced me to lock before I shopped for tools
North's Chapter 3 framing of request, kickoff, and scope stopped me from writing "add AI" as the request (North, 2023). The request is the freeze itself. Faced with signs of compromise or misconfiguration and no step list, a student has to decide what to check, when the evidence is enough, and what to change without taking the host down. Cathy Moore's "Will training help?" flowchart pointed the same direction: instruction earns its keep on the diagnostic sequence under fading support, not on another command reference (Moore, n.d.). The job aid, range self-check, published rubric, and AI-use protocol already carry part of that load.
Chapter 4 is where the work turned into a cognitive task analysis instead of a longer lab sheet. Host triage is the high-value task. The sequence I care about is mostly judgment: state the question, get a baseline, read processes and network state, check persistence and accounts, decide, contain, verify, then write a justification in 200 words or fewer. What actually drives the technology choice is when the work happens. On asynchronous nights there is no instructor in the room, and generative AI is often the only thing students can ask. I still grade the whole sequence across four scenarios. They need one evening that ends without feeling they broke the machine. A chat window that hands them the persistence path makes that gap worse. A partner that keeps them inside the triage order might not.
Two candidates, one primary
Module 3 wants an emerging-technology analysis. Students already work on live hosts in the Open Cyber Range, so I am not adding another simulation platform. I looked at ChatGPT Edu and at Canvas IgniteAI.
With Edu I can lock the model to the scenario brief, the one-page triage job aid, and the AI-use protocol. IgniteAI is the AI already built into Canvas, where students open announcements, files, and drops. That can help while they are studying course materials. It does not help much when they are alone on the range host at night and do not know what to check next.
I am going with ChatGPT Edu. Canvas still holds the brief, the protocol, the rubric, and the justification drop. In Scenario 1 the student attaches only the job aid and the brief, asks what category of evidence to gather next, and is told not to ask for the exact persistence path. The justification has to say which suggestion they used, which they ignored, and what the range actually showed. That is ATD Technology Literacy I can defend: the tool is part of the work, not a feature list (Association for Talent Development, n.d.). ISTE Visionary Planner work asks the same question I ask before I buy anything: does this solve a named need, or am I buying it because it is new (International Society for Technology in Education, 2018). IBSTPI still wants the boring answers on cost, access, and whether the environment can support the tool (Koszalka et al., 2013). I do not have those answers yet, so they stay labeled as assumptions until licensing and student devices are checked.
What this changes
Blog 1 left me with an AI Opportunities prompt and AI-free finals. Module 3 Week 1 makes the rule tighter. What is new here is not the range and not an LMS study chatbot. It is a constrained partner that can coach categories and cannot hand over the path. Bastani et al. (2025) already showed what unrestricted practice help can do to later performance; I would rather not rediscover that in my own section. The AI-use notes and the short justification become the window I get when I am not there. Gandini and Morselli (2026) found almost no student voice in challenge-based learning studies, and cybersecurity has the same hole. After Scenario 1 I need those voices before I claim the tool helps.
References
Association for Talent Development. (n.d.). Talent development capability model. https://td.org/capability-model
Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122(26), Article e2422633122. https://doi.org/10.1073/pnas.2422633122
Gandini, M. F., & Morselli, D. (2026). Challenge-based learning implementation in K-12 education: A systematic literature review. Frontiers in Education, 10, Article 1680347. https://doi.org/10.3389/feduc.2025.1680347
International Society for Technology in Education. (2018). ISTE standards for education leaders. https://iste.org/standards/education-leaders
Koszalka, T. A., Russ-Eft, D. F., & Reiser, R. (2013). Instructional designer competencies: The standards (4th ed.). Information Age.
Moore, C. (n.d.). Will training help? https://blog.cathy-moore.com/will-training-help/
North, C. (2023). Learning experience design essentials: Designing for users and impact. ATD Press.